The goal of this study is to identify risk and vulnerability mechanisms for alcohol and marijuana use-related problems among college students. Use of alcohol and marijuana is prevalent among college students. Use of these drugs increases the potential for a broad range of psychosocial problems. Research indicates both common and divergent risk mechanisms for substance use and use-related problems. This project aims to examine the association of four risk variables (impulsivity, affect lability, drug consumption, and perceived utility of substance use in achieving personal goals) with problems associated with marijuana and alcohol use among college students. These variables have all been linked to substance use problems in cross-sectional research. This study will explore relationships with a prospective longitudinal design. College students (n=800) will be recruited through e-mail and fliers on campus. Participants will be assessed at two time points separated by a period of six months. All questionnaires will be completed online in reserved computer labs in small group sessions under the supervision of a research assistant. All responses will be anonymous and participants will be identified only by a unique self-generated code. Analyses will focus on the subset of students who report using alcohol (85 percent) or marijuana (25 percent) during the six-month assessment window. Three types of relationships will be examined. The first relationship is an indirect path in which the association between the risk variable and use-related problems is mediated by use frequency. The second, is a direct path in which the risk variable contributes to use-related problems independently of use frequency. The significance of the direct and indirect paths will be determined by a structural equation model. Third, a vulnerability mechanism will be examined. In addition to a direct effect, impulsivity is hypothesized to be a vulnerability factor, moderating the relationship between use-related problems and drug consumption. Moderating effects will be determined by multiple regression models. Better understanding of the etiological paths of substance use problems may be useful for early identification of individuals at risk for substance use problems and may inform the development of effective interventions.